Question: Professor Kreimeyer, your doctoral thesis focused on complexity management. Has product development become more complex since then, and how can that complexity be measured?

Kreimeyer: I believe it has. To use a buzzword, cyber-physical systems have made the world more complex. More disciplines are involved, there are more human factors to consider, services have become part of the equation, and so on. As a result, we have a highly dynamic network of many different perspectives, and that is also how I understand complexity. In my doctoral thesis, I tried to develop metrics for measuring this interconnectedness.

Question: What is the biggest driver of complexity today: software, sustainability, excessive regulation, globalization, or something else entirely?

Kreimeyer: That is difficult to say because complexity has so many layers. Each of these layers is complex enough in its own right. The real challenge is, first, to coordinate all of them, which is what we now call systems integration. Second, it involves finding the right balance. Being well integrated does not automatically mean being well balanced. You can still develop a product that performs poorly in terms of sustainability while striking all the other compromises well.

Question: In the past, products primarily had to work. Today, they are also expected to be sustainable, circular, and as low-carbon as possible. Aren't we overwhelming product developers with all these conflicting objectives?

Kreimeyer: I don't think so. Of course, development is becoming more complex and therefore more demanding, but well-trained engineers are already capable of resolving such conflicts. The greater difficulty is doing so under time and resource pressure. I do expect there to be a natural limit, however, because you cannot keep adding more layers and still expect to become faster. At the same time, we can see what some companies in China are doing. It shows that AI and local engineering can significantly accelerate development.

Question: Are Chinese companies developing products faster because their methods are better, or simply because they work longer hours?

Kreimeyer: Our Western European engineering culture is more risk averse, which is also related to the large number of regulations. We plan everything down to the last detail and validate it 150 percent. That often slows development because it simply takes more iterations. My hypothesis is that other engineering cultures are better at this. The same is true of the American companies. Just look at how many updates Tesla needed in the early days. Many of them only came after the cars had gone on sale.

Question: Do we need to rethink our development methods in order to create sustainable products?

Kreimeyer: Perhaps we do not necessarily need to rethink them, but we do need to expand them considerably. Here in Stuttgart, we are doing quite a bit of work on what a system architecture for sustainability should look like so that materials, components, or even entire parts of a system can be reused in future products. It is an exciting topic because I have to think about what the next product generation will look like while I am still developing the current one. But I would say that this kind of advance planning fits our Western European engineering culture very well. Sustainable products that remain on the market longer and eventually have to be dealt with again naturally increase complexity considerably because, for example, the logistics chain also has to be taken into account.

Question: Model-Based Systems Engineering (MBSE) is intended to make complexity easier to manage, but it is itself complex. How can companies get started without becoming overwhelmed?

Kreimeyer: MBSE is complex because it requires a higher level of abstraction, but it offers many advantages. Especially for a midsized company, the best way to get started is simply to have someone take a two-week training course and model a single part of a product to see how it works. I think you quickly realize that it is not actually that complicated. In some areas, however, you have to be more disciplined than in the past because you work more closely with other disciplines.

Question: Everyone is currently talking about what AI can do. The more interesting question is what it is doing to engineers. What is your view?

Kreimeyer: Let me look at it from a university perspective. AI is developing so quickly that we cannot establish rules for studying and teaching as fast as the models improve. In general, we see that students often find it very difficult to assess what AI produces. We also have good students who compare and question the results. But I believe that, both during their education and later in their careers, it will become more difficult for engineers to determine what corresponds to reality. AI is dramatically changing the way engineers work, but I am not very concerned that it will take away jobs on a large scale. We will become faster, and I see that as a positive development.

Question: Can you illustrate that with an example?

Kreimeyer: Today, we might run ten simulations and build one prototype to check whether the simulation is correct. In the future, we will probably conduct an AI-based parameter study or use a surrogate model, but I hope we will still build a prototype so that we can critically assess the AI results. When you work with an FEM model, you develop an intuitive sense of what works and what does not because you can see how the model responds. AI does not give you that intuition. The consequence is that engineering teams working with AI somehow have to calibrate themselves. We do that by still making our students calculate things by hand or build an FEM model. Showing them how to use AI intelligently is one of our most important tasks. To be honest, the students often do not particularly enjoy it.

Question: Engineers have less and less time to develop products because they have to manage more and more information. Can AI free them from this burden?

Kreimeyer: I don't like the word 'burden' because I believe that engineers are primarily information hubs within a company. They have to balance all the trade-offs we talked about earlier in terms of manufacturability, cost optimization, usability, sustainability, and so on. Being able to communicate well is therefore the key to engineering. I do believe, however, that AI models can help us evaluate and reconcile the many streams of information that converge on engineers and sometimes contradict one another.

Question: What skills will good engineers still need ten years from now that no AI can replace?

Kreimeyer: Essentially what I just said: good communication skills and the ability to talk to an increasing number of disciplines. The ability to think from the customer's perspective will also become even more important in view of cyber-physical systems and new business models. And the third skill, when we talk about MBSE, is certainly the ability to think abstractly. I became a mechanical engineer because I was good at visualizing things in three dimensions. Today, I am miles away from that.

Question: Do students learn these communication skills at university today?

Kreimeyer: This is one area where I would criticize universities because we do far too little when it comes to soft skills. I am talking specifically about presenting, facilitating, and leading without formal authority. The best way to learn these things is by doing and by working in teams. That is why I recommend that my students join Formula Student teams, which exist at every university, get involved in clubs and associations, or gain experience through in